That was the subject of “News albo fake news” on Polish Radio Programme 3 (Trójka), which featured a conversation with Sylwia Adamczyk, an expert at NASK’s Department for the Protection of Information Security in Cyberspace, and Dr Nikodem Sarna, a digital marketing expert at Kozminski University. The guests talked about what bots are, how they work and how they shape our sense of reality online.
Dr Nikodem Sarna explains that a bot is software that automatically performs actions previously carried out by humans. “A bot is a piece of software, that is, a program that carries out various actions. Before bots came along, these actions were reserved for people; now they are performed online automatically, under someone’s supervision but without human involvement,” the expert says. Bots can imitate human behaviour and mislead users, but they are also used for useful purposes — passing on information or doing work that would be too exhausting for a human or would take almost for ever. The very existence of search engines that return more or less relevant results on request is an example of the positive use of bots replacing human labour. At the same time, Sarna stresses, there are a great many less cheerful and downright dangerous applications.
It is on these threats that Sylwia Adamczyk focuses when she talks about the role of bots in disinformation. According to her, bots are used on a mass scale to imitate human activity and to create artificial information noise around false claims. “They are often used to imitate human behaviour and to generate activity online around false narratives,” the expert explains. So-called bot networks or bot farms can number thousands or even tens of thousands of accounts, which comment on posts containing fake claims and interact with one another and with real users, creating the impression of mass support. A user who sees a large number of comments under a false post may conclude that the information is true, because “so many people can’t be wrong”.
In such situations, Adamczyk notes, a cognitive bias known as social proof kicks in. “When we see that a claim has a lot of support, we think: ‘Well, there must be something in it, it’s probably true,’” she says. That is why disinformation campaigns so often use bots on a large scale.
At the same time, the expert stresses that bots are not independent actors. “There is always someone behind these bots — someone directing what they do,” she emphasises. Bot activity is especially effective when it appeals to emotions, because emotional content spreads fastest online. People themselves unwittingly help this along by reacting to anger, fear or outrage, but platform algorithms play their part too. They are designed not to keep users well informed but to keep them on the platforms for as long as possible — time that can be monetised. That is why algorithms so often promote emotional, frequently negative content, which is exactly what disinformation “feeds” on, creating a vicious circle.
Nikodem Sarna points out that statistics confirm the scale of bot activity. According to Cloudflare, in 2025 only about 43.5 per cent of all internet traffic was generated by humans, while more than half came from bots. A large share of this “non-human” traffic is accounted for by so-called “good” bots, such as those that index content for search engines or collect data for AI tools. The rest, however, are “bad” bots, which are openly harmful or even criminal.
These include bots that steal transaction data from online forms, generate artificial traffic to knock servers offline, or automatically add huge numbers of items to shopping baskets, creating the illusion that they are out of stock. Such actions, known as DDoS attacks, are often a problem for ticketing services and other online platforms. Alongside this there is a “grey zone” of bot influence — the question of how the sheer volume of bot traffic affects the quality of the internet experience for ordinary people, their access to services and overall convenience.
In this context, talk has returned to the “dead internet” theory, which emerged as far back as 2016 and was at first regarded as a conspiracy theory. It holds that there is less and less human content online, while bot activity keeps growing. “Certain elements of this theory are starting to come true,” Sylwia Adamczyk observes, adding that bots no longer just generate content but actively interact with one another. This raises a further question: what will large language models “feed” on in the future, and will they start learning mainly from content created by other algorithms?
Nikodem Sarna also points out that in today’s world information increasingly ceases to be an advantage and begins to complicate our lives. Although almost everyone has access to the internet, that access is paradoxical, since it does not necessarily lead to knowledge and understanding of the world. Instead, users are plunged into an endless stream of stimuli. In such an environment, companies and media that try to act honestly find it ever harder to build relationships with their audiences and to create convenient, reliable online services.
Sylwia Adamczyk adds that algorithmic personalisation of content locks users into so-called information or filter bubbles. In their news feeds people mostly see what is meant to please them and grab their attention, rather than what broadens their outlook or challenges their beliefs. “What we consume as ‘interesting’ is often just empty calories,” the expert remarks, stressing that this is also a danger for the media, which sometimes fall into the trap of fakes themselves.
A particularly serious problem is the penetration of deepfakes into traditional media, from which audiences expect verified and reliable information. Adamczyk gives examples of deepfakes with obvious artefacts being aired on television — fakes that could easily have been spotted, but weren’t. This shows that the media face a difficult task: learning to work with new technologies without undermining their audiences’ trust.
Ultimately, as the participants conclude, there are no simple solutions here. New technologies are a challenge for experts, journalists and ordinary users alike. The media are staffed by the same people, for whom these tools are also new, so we will have to work it all out together, gradually learning to tell reality from imitation.
Dr. Myroslav Trofymuk
Born in Lviv. Designer, director, musician, researcher, founder of ZEMLA STUDIO. Master of classical philology and digital marketing, PhD in the history of Ukrainian literature. Member of the National Union of Journalists of Ukraine. Since 2022 he has authored the CENTRUM MUNDI audio show for Polish Radio for Ukraine, continuing the same research into the contemporary music scene in Ukraine and Poland.
Full bio →
